What is market microstructure?
Market microstructure is the study of how trading actually happens, from the way orders are placed, matched and reported to the effect those mechanics have on prices, spreads and costs. Where most of finance treats a price as given, microstructure asks how it was formed, trade by trade and quote by quote.
Key facts
| What it studies | How orders are placed, matched and reported, and the effect on prices, spreads and costs |
|---|---|
| Typical questions | Who supplied liquidity, what a trade cost against the quotes and how much volume happened away from the exchanges |
| Data needed | Trades and quotes in time order, finer than a price chart |
| Built on | Tick data in which every quote and trade carries a timestamp |
| In algoseek data | TradeAtBid, TradeAtMid, TradeAtAsk and TimeWeightSpread in US Equities Trade and Quote Extended Minute Bar, and tick data in US Equities Trade and Quote |
What questions does market microstructure deal with?
The questions it deals with are concrete. Who supplied liquidity and who took it, what a trade cost relative to the quotes standing at the time, how much of the volume happened away from the exchanges and how quickly prices absorbed the information in an order. For US stocks the answers are shaped by a market that the SEC has described as dispersed and complex, with trading split across many exchanges and off-exchange venues that are linked by rules such as Regulation NMS.
What does market microstructure look like in algoseek data?
Answering any of those questions takes data finer than a price chart. A single one-minute bar from algoseek’s US Equities Trade and Quote Extended Minute Bar dataset (US1034), Apple at 11:00 on October 9, 2026, already shows several layers of it:
US Equities Trade and Quote Extended Minute Bar · AAPL, 11:00 on October 9, 2026
| Field | Value |
|---|---|
| OpenBidPrice / OpenAskPrice | 334.06 / 334.1 |
| CloseBidPrice / CloseAskPrice | 334.25 / 334.28 |
| TimeWeightSpread | 0.05012 |
| NBBOQuoteCount | 2756 |
| TotalTrades | 1476 |
| OddLotTradeCount | 826 |
| Volume (exchanges) | 33183 |
| FinraVolume (off-exchange) | 56808 |
| TradeAtBid | 13810 |
| TradeAtMid | 32238 |
| TradeAtAsk | 26391 |
| TradeAtBidCount | 281 |
| TradeAtAskCount | 356 |
Read together, the rows describe the minute in a way four prices cannot. The bar counts 2,756 national best bid and offer quote records against 1,476 trades, and 826 of those trades were odd lots.
Off-exchange venues reported 63% of the minute’s 89,991 shares, and 36% of all shares changed hands at the exact midpoint, inside a spread that averaged five cents over the minute.
Where did each trade print against the quotes?
The TradeAt fields carry the classic microstructure measurement, where each trade printed relative to the prevailing quotes. Nearly twice as many shares traded at or above the offer as at or below the bid in this minute, 26,391 against 13,810, and over the same minute the midpoint rose by 18.5 cents.
Counting trades instead of shares narrows the gap to 356 at the offer against 281 at the bid. Trades at the offer are often read as started by buyers and those at the bid as started by sellers, but the data records only where each trade printed against the quotes, not which side initiated it.
What gets lost when market data is aggregated?
For anyone working with market data, microstructure is mostly a matter of what gets lost in aggregation. A daily bar has no spread in it and a minute bar loses the order of trades and quotes, while a single volume figure says nothing about where the shares changed hands. Execution analysis, transaction cost models and short-horizon research all need those details back, which is why they are built on tick data such as US Equities Trade and Quote (US1032), where every quote and trade carries a nanosecond timestamp.
Frequently asked questions
-
What data does market microstructure research need?
+
It needs data finer than a price chart. Execution analysis, transaction cost models and short-horizon research are built on tick data, where every quote and trade carries its own timestamp.
-
What is the bid-ask spread?
+
It is the gap between the best bid and the best offer. In the Apple minute shown above it averaged five cents, and 36% of all shares changed hands at the exact midpoint inside it.
-
How are trades classified as buys or sells?
+
Trades at the offer are often read as started by buyers and those at the bid as started by sellers. The data records only where each trade printed against the quotes, not which side initiated it.